cdm AC Configuration

This article is for probe versions 5.6 or later.
uimpga-ga
cdm_AC
This article describes the configuration concepts and procedures to set up the CPU, Disk, Memory Performance Monitoring (cdm) probe. You can configure the probe to monitor the CPU, disk, and memory of the system on which the probe is deployed. The probe automatically identifies components and allows you to specify QoS parameters and thresholds for alarms.
 
This article is for probe versions 5.6 or later.
 
The following diagram outlines the process to configure the probe.
Configuring CDM AC
Configuring CDM AC
 
Contents
 
 
 
2
 
 
Verify Prerequisites
Verify that required hardware and software is available and any installation and upgrade considerations are met before you configure the probe. For more information, see cdm (CPU, Disk, Memory Performance Monitoring) Release Notes.
Configure General Properties
You can configure the logging properties of the probe. You can also specify global QoS and alarm source and target parameters.
 
Follow these steps:
 
  1. Select the 
    cdm
     node.
    The
     Probe Information 
    section provides information about the probe name, probe version, start time of the probe, and the probe vendor. 
  2. Under the 
    General Configuration
     section, update the following information:
    •  
      Log level
      :
       
      specifies the level of details that are written to the log file. You can select the following log levels:
      • 0 - Logs only severe information (default)
      • 1 - Logs error information
      • 2 - Logs warning information
      • 3 - Logs general information
      • 4 - Logs debugging information
      • 5 - Logs tracing/low-level debugging information
      Log as little as possible during normal operation to minimize disk consumption, and increase the amount of detail when debugging.
    •  
      Log size (KB)
      : specifies a maximum size of the probe log file.  Upon reaching this size, the contents of the file are cleared.
      Default: 100 KB
    •  
      Send alarm on each sample: 
      allows you to generate an alarm on each sample, if selected. If not selected, the probe waits for the applicable number of samples before sending the alarm. For example, if the interval is specified as 1 minute and the number of samples is specified as 2, and if this option is:
      •  
        Unchecked:
         the probe generates the first alarm in 2 minutes and each respective alarm in 1 minute interval.
      •  
        Checked:
         the probe generates the first alarm in 1 minute and each respective alarm in 1 minute interval.
        Default: Selected
         The sample collected at the start of the probe is considered as the first sample. The sample count is cleared on probe deactivation.
    •  
      Send short name for QoS source: 
      allows you to send only the host name as the source in QoS messages. If not selected, sends the FQDN (Fully Qualified Domain Name).
       If the 
      Set QoS source to robot name
       option is selected in the controller, robot name is used as target.
    •  
      Allow QoS source as target: 
      allows you to use the QoS source as the target name. QoS messages can use the host name as their target, by default. If selected, the target name is changed to the QoS source name.
    •  
      (Linux, AIX and Solaris platforms only) Monitor iostat:
       enables the iostat monitoring of the host system devices.
    •  
      (Linux, Solaris, AIX and HP-UX platforms only) Count Buffer-Cache as Used Memory
      : counts the buffer and cache memory as used memory while monitoring the physical and system memory utilization. If not selected, the buffer and cache memory is counted as free memory.
    •  
      (Linux, Solaris, AIX and HP-UX platforms only) Calculate Load Average per Processor: 
      allows you to calculate the load average per processor. For all Unix-based systems, the system load measures the computational work performed. This means that if your system has a load of 4, then four running processes are either using or waiting for the CPU. Load average refers to the average of the computer load over several periods of time.
      Default: Selected
  3. Click 
    Save.
     
Configure Cluster Alarm and QoS Source
The probe automatically detects and displays the list of virtual groups belonging to the cluster.  You can configure the alarm and QoS source for each virtual group as one of the following parameters:
  •  
    <cluster ip>:
     The IP address of the cluster.
  •  
    <cluster name>:
     The name of the cluster.
  •  
    <cluster name>.<group name>:
     A combination of the cluster name and the group name
 The 
Cluster
 tab is available only when the monitored system is part of a cluster and the 
cluster
 probe is deployed on the robot.
 
Follow these steps:
 
  1. Navigate to the 
    cdm
     node >
     Cluster
     section.
    The probe displays all the cluster groups of the monitored system. 
  2. Select the required cluster from the list. The probe displays the name and IP address of the cluster and the virtual group that the monitored system is a part of.
  3. Update the following information to configure the alarm and QoS source:
    •  
      Alarm Source:
       specifies the source to be used for alarms.
      Default: <cluster ip>
    •  
      QoS Source:
       specifies the source to be used for QoS messages.
      Default: <cluster ip> 
  4. Click 
    Save
    .
Configure Host Monitors
You can configure the following monitors for the host where the probe is deployed:
  • Host uptime
  • Host reboot
 
Follow these steps:
 
  1. Navigate to the 
    cdm >
     
     
    Host name
     
     node 
    > Monitors
     section.
  2. Select the 
    Computer Uptime Hourly
     monitor from the list.
  3. Select 
    Publish Data
     to generate hourly QoS messages for the computer uptime (in seconds).
  4. Select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  5. Select the 
    System Reboot
     monitor from the list.
  6. Select 
    Publish Alarms
     to generate an alarm when the probe detects a system reboot on the host.
  7. Select the alarm message from 
    Alarm Message for Detected Reboot
     that the probe generates when the host system reboots.
  8. Click 
    Save
    .
Configure Alarm and QoS Properties
You can configure the following alarm and QoS properties of the probe:
 
 
3
 
3
 
 
For any issue related to incorrect QoS data or command output, see cdm Troubleshooting.
Configure Default Disk Monitoring Parameters
You can configure the default monitoring parameters for all disks that are monitored on the host.
 
Follow these steps:
 
  1. Navigate to the 
    cdm > 
    Host name
     > Disks
     node. The probe displays all the disks for the host as 
     
    Disk name
     
     nodes.
  2. Under the 
    Disks
     node 
    > Disk Configuration
     section, update the following information to configure monitoring parameters:
    Reduce the intervals and number of samples to generate alarms and QoS frequently. Smaller values can also increase the system load.
    •  
      Interval: 
      specifies the interval (in minutes) after which the monitoring information is obtained.
    •  
      Samples: 
      specifies the number of samples the probe stores to calculate values to compare with thresholds.
      : Set the sample value to 0 to use the default sample value for QoS messages for 
      disks
      .
    •  
      QoS Interval (Multiple of 'Interval')
      : specifies the monitoring interval between each sample collection for QoS information. For example, If the interval is set to 5 minutes and number of samples is set to 5, the average CPU utilization for the last 25 minutes is displayed in the QoS message.
       If “the value is greater than 1, the probe only calculates the average data for CPU monitoring. For disk, memory, and paging QoS messages, the probe generates the QoS messages with the current value. The probe only uses the QoS Interval to reduce the frequency of QoS messages.
    •  
      Ignore Filesystems: 
      defines the file system to be excluded from monitoring. For example, specify the regular expression 
      C:\\
       to exclude the C drive of the system from monitoring.
       On UNIX platforms, use the regular expression 
      (/\)
       to exclude the root directory 
      (/)
       from monitoring.
    •  
      Timeout(In Seconds): 
      specifies the maximum time for the probe to collect monitoring information. For example, timeout for disk fail or crash scenarios in stale filesystems allow you to prevent the probe from going into a pending state. You can specify the default timeout of 5 seconds to retrieve the disk statistics.
       CA recommends a value of 10 seconds when the monitored system has high CPU load.
    •  
      Filesystem Type Filter: 
      specifies the type of the filesystem to be monitored using regular expressions.
      • If we specify RegEx as *, then all filesystems are enabled for monitoring. Or you can also specify ext* to allow monitoring of filesystems with "ext". For example, ext4 or ext5.
      • If we specify a negative RegEx, then specified filesystems are excluded from monitoring. For example, if we do not want to monitor ext4 filesystem, use /^(?!ext4)/
      • If this field is blank, no filesystem is enabled for monitoring. (default)
  3. Under the 
    Disk Missing Defaults
     section, update the following information to configure default parameters for missing disks:
    •  
      Disk Missing Alarm: 
      enables the probe to generate an alarm if a monitored disk is missing from the host.
    •  
      Alarm Message:
       specifies the alarm message to generate if a monitored disk is missing from the host.
  4. Under the 
    Disk Usage Defaults
     section, update the following information to configure default parameters for disk usage:
    •  
      Publishing Alarm Based on:
       select from the following options to specify the measuring criteria for disk usage:
      •  
        MB: 
        enables the probe to generate disk usage alarms in megabytes.
      •  
        percent : 
        enables the probe to generate disk usage alarms in percentage.
      Default: percent
    •  
      Enable High Threshold: 
      enables the probe to generate an alarm if the free disk space is lower than the specified value. CA recommends you to specify a higher severity alarm for High threshold.
      •  
        Threshold: 
        specifies the minimum free disk space. The probe generates the alarm if the disk usage is lower than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the free disk space is lower than the specified value.
        Default severity: Error
    •  
      Enable Low Threshold: 
      enables the probe to generate an alarm if the free disk space is lower than the specified value. CA recommends you to specify a lower severity alarm for Low threshold.
      •  
        Threshold: 
        specifies the minimum free disk space. The probe generates the alarm if the disk usage is lower than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the free disk space is lower than the specified value.
        Default severity: Warning
    •  
      (Disk Usage %) Publish Data: 
      enables the probe to generate QoS messages on disk usage (in percentage).
    •  
      (Disk Usage MB) Publish Data: 
      enables the probe to generate QoS messages on disk usage (in megabytes).
  5. Under the 
    Inode Usage Defaults
     section, update the following information to configure default parameters for inode usage:
    •  
      Inode Usage Alarm Based on Threshold for:
       select from the following options to specify the measuring criteria for inode usage:
      •  
        MB: 
        enables the probe to generate inode usage alarms in megabytes.
      •  
        percent : 
        enables the probe to generate inode usage alarms in percentage.
      Default: percent
    •  
      Enable High Threshold: 
      enables the probe to generate an alarm if the free inode space is lower than the specified value. CA recommends you to specify a higher severity alarm for High threshold.
      •  
        Threshold: 
        specifies the minimum free inode space. The probe generates the alarm if the free inode space is lower than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the free inode space is lower than the specified value.
        Default severity: Error
    •  
      Enable Low Threshold: 
      enables the probe to generate an alarm if the free inode space is lower than the specified value. CA recommends you to specify a lower severity alarm for Low threshold.
      •  
        Threshold: 
        specifies the minimum free inode space. The probe generates the alarm if the free inode space is lower than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the free inode space is lower than the specified value.
        Default severity: Warning
    •  
      (Inode Usage %) Publish Data: 
      enables the probe to generate QoS messages on inode usage (in percentage).
    •  
      (Inode Usage MB) Publish Data: 
      enables the probe to generate QoS messages on inode usage (in megabytes).
  6. Under the 
    Disk Usage Change Defaults
     section, update the following information to configure default parameters for change in disk usage. You can configure these parameters to generate QoS in MB or Percentage:
    •  
      Send Alarm Based on Free Space:
       allows you to specify whether alarms are generated in MB or % for free space.
    •  
      Type of change:
       allows you to specify whether alarms are generated on increase, decrease, or both increase and decrease in disk usage. 
    •  
      Change Calculation:
       allows you to select one of the following options to configure disk usage change calculation:
      •  
        Summarized over all samples:
         enables the probe to use the difference between the latest sample and the first sample.
        : There can be discrepancies between the values in QoS and values in alarms when the 
        Summarized over all samples 
        option is selected. This is because the QoS are generated on every interval and alarms are generated based on the selection of the option 
        Summarized over all samples
        .
      •  
        Between each sample:
         The change in disk usage will be calculated after each sample is collected.
    •  
      Enable High Threshold: 
      enables the probe to generate an alarm if the change in disk usage is greater than the specified value. CA recommends you to specify a higher severity alarm for High threshold.
      •  
        Threshold: 
        specifies the maximum change in disk usage. The probe generates a higher severity alarm if the change in disk  usage is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the change in disk usage is greater than the specified value.
        Default severity: Error
    •  
      Enable Low Threshold: 
      enables the probe to generate an alarm if the change in disk usage is greater than the specified value. CA recommends you to specify a lower severity alarm for Low threshold.
      •  
        Threshold: 
        specifies the maximum change in disk usage. The probe generates a lower severity alarm if the change in disk usage is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the change in disk usage is greater than the specified value.
        Default severity: Warning
    •  
      Publishing Data: 
      enables the probe to generate QoS messages on change in disk usage.
  7.  
    (Windows, Linux and AIX platforms)
     Under the 
    Monitors
     section, select the monitor from the list and then 
    Publish Data
     to generate QoS messages for disk throughput.
  8.  
    (Windows, Linux and AIX platforms)
     Update the following information to configure the thresholds for the alarm message:
    •  
      Enable High Threshold: 
      enables the probe to generate an alarm if the disk throughput is greater than the specified value. CA recommends you to specify a higher severity alarm for High threshold.
      •  
        Threshold: 
        specifies the maximum disk throughput. The probe generates a higher severity alarm if the disk throughput is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the disk throughput is greater than the specified value.
        Default severity: Error
    •  
      Enable Low Threshold: 
      enables the probe to generate an alarm if the disk throughput is greater than the specified value. CA recommends you to specify a lower severity alarm for Low threshold.
      •  
        Threshold: 
        specifies the maximum disk throughput. The probe generates a lower severity alarm if the disk throughput is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the disk throughput is greater than the specified value.
        Default severity: Warning
  9. Select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  10. Click 
    Save
    .
Configure Alarms and QoS for Disk Usage
You can configure the probe for local, cluster, and shared (network) disk monitoring. You can configure thresholds for alarms and QoS.for inodes. When monitoring cluster or shared disks (such as NFS mounts) over low-performance or over-utilized lines, the response time can be slow.
 The NFS mounts monitored in the cdm probe point to servers. These servers also appear during discovery in USM.
You can generate QoS messages and alarms for the following disk performance categories:
  • Disk usage on local, clustered, and shared disks (in percentage or megabytes). Alarms are generated for free disk space and QoS are generated for disk usage.
  • Change in disk usage on local and clustered disks (in megabytes)
The probe uses the mount entries as in /proc/mounts file in Linux to display the file system type of devices that are remounted to a different location.
 
Follow these steps:
 
  1. Navigate to the 
    cdm > 
    Host name
     > Disks > 
    Disk name
     
     node.
  2.  
    (Network Shared Disks)
     Select the 
     
    Disk name
     
     node for a shared disk and select 
    Enable Space Monitoring
     to configure disk usage on shared (network) disks.
  3.  
    (Local and Clustered Disks)
     Under the 
    Monitors
     section, select the monitor from the list. 
  4.  
    (Local and Clustered Disks)
     Select 
    Publish Data
     and 
    Publish Alarms
     fields to retrieve QoS data and generate alarm messages, respectively.
    : You can generate alarms only for the 
    Disk Missing
     monitor. Select the alarm message from the 
    Alarm Message
     drop-down list. Disk size alarms and QoS are supported only on the Windows, Linux, and AIX platforms.
  5. Select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  6. Navigate to the 
    Disk Usage
     node > 
    Alarm Thresholds
     section and specify the values in the fields.
  7.  
    (Optional)
     Navigate to the 
    Inode Usage
     node > 
    Alarm Thresholds
     section and specify the values in the fields.
  8. Navigate to the 
    Disk Usage Change
     node > 
    Monitors
     section and specify the values in the fields.
  9. Select the required monitor from the nodes and then 
    Publish Data
     to generate QoS messages for the monitor.
  10. Select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  11. Click 
    Save
    .
(Windows) Monitor Shared Disk Availability
You can monitor the availability of shared (network) disks and generate alarm and QoS messages.
 
Follow these steps:
 
  1. Click the Options (icon) next to Disks node and select 
    Add New Share
    .
  2. Update the following fields in the 
    Add New Share
     dialog:
    •  
      Share: 
      displays the path of the shared disk.
    •  
      User:
       specifies the username of the account that can access the shared disk.
    •  
      Password:
       specifies the password for the specified 
      User
      .
    •  
      Alarm Message:
       specifies the alarm that the probe generates when connection to the shared disk fails.
  3. Select 
    Enable Folder Availability Monitoring
     to enable the probe to generate QoS messages for the shared folder availability.
  4. Click 
    Submit
     to close the dialog.
  5. Select the 
     
    Disk name
     
     node for the shared disk and select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  6. Click 
    Save
    .
Configure Memory Monitoring
You can configure the monitoring parameters to monitor the following memory categories:
  • Paging memory
  • Total memory
  •  
    Swap memory:
     A swap memory is a reserved space on hard drive which is used by the system when the physical memory (RAM) is full. However, the swap memory is not a replacement of the physical memory due to lower data access rate. The probe calculates the swap memory similar to the swap -l command of Solaris. However, the probe use pages instead of blocks. You can compare the swap memory information of the probe and the swap -l command by using the following formula:
    Swap Memory (calculated by probe) in MB = (Blocks returned by the swap -l command * 512)/ (1024*1024)
  •  
    Physical memory:
     You can monitor the total physical memory and memory utilization of the system to generate QoS data and configure threshold to generate alarms. The memory utilization is directly related to the system performance. If the memory utilization is high, the system performance goes down and the application response time increases. The increased response time of critical business applications can adversely impact the user interaction. Monitoring the system memory lets you helps diagnosing the issue, for example, identifying the closing of unwanted applications. You can also consider system upgrade when the memory utilization is consistently high. 
  •  
    (From version 6.10 on AIX only)
     With Active Memory Expansion (AME) mode enabled, you can generate QoS messages for the following monitors:
    •  
      Memory > Physical Memory
      •  
        Memory Dxm
         monitors the memory deficit in megabytes. At times, a logical partition (LPAR) cannot be configured with the provided memory expansion factor as it is too large, and the workload in the LPAR does not compress well. Thus, a memory deficit is created. You can also generate alarms for this monitor.
      •  
        Memory Xphysc
         monitors the number of physical processors used for AME.
    •  
      Memory > Total Memory
      •  
        Memory Expansion Factor
         monitors the target memory expansion factor that you have configured in your system, to increase the total physical memory of an LPAR in your system.
      •  
        Total Memory
         monitors the total memory size of an LPAR in megabytes.
         This monitor also generates QoS messages on a non-AME mode.
 
Follow these steps:
 
  1. Navigate to the 
    cdm > 
    Host name
     > Memory 
    node.
  2. Under the 
    Memory Configuration
     section, update the following information to configure monitoring parameters:
    Reduce the intervals and number of samples to generate alarms and QoS frequently. Smaller values can also increase the system load.
    •  
      Interval (minutes): 
      specifies the interval (in minutes) after which the monitoring information is obtained.
       On Solaris Global Zone platforms, the probe executes the command 
       
      echo ::memstat | mdb -k
       
       to retrieve the buffer and cache memory values. Ensure that the specified alarm and QoS intervals in the probe are greater than 
      mdb
       command response time. Otherwise, perform one of the following options in the Raw Configuration interface:
      • Set the value of 
        mem_buffer_timeout
         key to a desirable time lesser than the profile execution interval time. The default value for this key is 
        25 sec. 
         
      • Set the 
        mem_buffer_used
         key to 
        Yes 
        so that the probe does not execute this command. For more information about this key, see cdm Advanced Configuration.
    •  
      Samples: 
      specifies the number of samples the probe stores to calculate values to compare with thresholds.
      : Set the sample value to 0 to use the default sample value for QoS messages.
    •  
      QoS Interval (Multiple of 'Interval')
      : specifies the monitoring interval between each sample collection for QoS information. For example, If the interval is set to 5 minutes and number of samples is set to 5, the average CPU utilization for the last 25 minutes is displayed in the QoS message.
       If “the value is greater than 1, the probe only calculates the average data for CPU monitoring. For disk, memory, and paging QoS messages, the probe generates the QoS messages with the current value. The probe only uses the QoS Interval to reduce the frequency of QoS messages.
  3. Select 
    Set QoS target as 'Memory' 
    to specify the QoS target for memory and paging as 
    Memory
    . The following SQL scripts demonstrate how to update old data in the database when the QoS target as 
    Memory
     is changed:
    • Execute the following SQL query in the UIM database to view the rows to be updated:
      SELECT * FROM dbo.s_qos_data
      WHERE probe LIKE 'cdm'
      AND (qos LIKE'QOS_MEMORY_PERC_USAGE'
      OR qos LIKE 'QOS_MEMORY_PAGING_PGPS'
      OR qos LIKE 'QOS_MEMORY_PAGING'
      OR qos LIKE 'QOS_MEMORY_PHYSICAL'
      OR qos LIKE 'QOS_MEMORY_PHYSICAL_PERC'
      OR qos LIKE 'QOS_MEMORY_SWAP'
      OR qos LIKE 'QOS_MEMORY_SWAP_PERC'
      OR qos LIKE 'QOS_PHYSICAL_MEMORY_TOTAL')
    • Execute the following SQL query in the UIM database to update the table for the new target. 
      Target
       is the new QoS target to be set. You can configure the value.
      Declare @Target varchar(100)
      SELECT @Target = 'Memory'
      UPDATE dbo.s_qos_data
      SET target=@Target
      WHERE probe LIKE 'cdm'
      AND (qos LIKE'QOS_MEMORY_PERC_USAGE'
      OR qos LIKE 'QOS_MEMORY_PAGING_PGPS'
      OR qos LIKE 'QOS_MEMORY_PAGING'
      OR qos LIKE 'QOS_MEMORY_PHYSICAL'
      OR qos LIKE 'QOS_MEMORY_PHYSICAL_PERC'
      OR qos LIKE 'QOS_MEMORY_SWAP'
      OR qos LIKE 'QOS_MEMORY_SWAP_PERC'
      OR qos LIKE 'QOS_PHYSICAL_MEMORY_TOTAL')
  4.  
    Top Memory consuming processes in alarm: 
    specifies the number of top memory consuming processes that are included in memory usage alarms.
    Default: 5
  5.  
    (Paging only)
     Navigate to the 
    Memory Paging
     node > 
    Alarm Thresholds
     section.
  6.  
    (Paging only)
     Select from the following values in the 
    Send Alarm Based on Threshold for
     drop-down list:
    •  
      KB/second: 
      enables you to generate QoS messages for the amount of paging memory in Kbytes/second.
    •  
      pages/second: 
      enables you to generate QoS messages for the amount of paging memory in pages per second.
  7. Select the required monitor from the nodes and then 
    Publish Data
     to generate QoS messages for the monitor.
  8. Update the following information to configure the thresholds for the alarm message:
    •  
      Enable High Threshold: 
      enables the probe to generate an alarm if the monitored value is greater than the specified value. CA recommends you to specify a higher severity alarm for High threshold.
      •  
        Threshold: 
        specifies the maximum value. The probe generates a higher severity alarm if the monitored value is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the monitored value is greater than the specified value.
        Default severity: Error
    •  
      Enable Low Threshold: 
      enables the probe to generate an alarm if the monitored value is greater than the specified value. CA recommends you to specify a lower severity alarm for Low threshold.
      •  
        Threshold: 
        specifies the maximum value. The probe generates a lower severity alarm if the monitored value is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the monitored value is greater than the specified value.
        Default severity: Warning
  9. Select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  10. Click 
    Save
    .
(Windows, Linux, and AIX platforms) Configure Network Traffic Monitoring
You can configure monitors for the outbound and inbound traffic of your system Network Interface Card (NIC). The NIC monitoring lets you analyze the network bandwidth that is being utilized which can impact the overall network performance. For example, your NIC capacity is 100 MBPS and aggregated traffic is more than 90 MBPS then it can slow down the data transfer rate. This monitoring helps you take preventive actions before the network goes down. For example, upgrade your NIC or install more NICs and implement the load-balancing solution.
The probe monitors only physical NICs of system and sum up the metric values when multiple NICs are installed on the monitored system.
This node lets you monitor the following network metrics:
  • Inbound Traffic: Monitors the traffic coming from LAN or a public network to the monitored system in bytes per second.
  • Outbound Traffic: Monitors the traffic going from the monitored system to LAN or a public network in bytes per second.
  • Aggregated Traffic: Monitors both inbound and traffic in bytes per second.
 
(From version 6.10) Note:
 The network traffic monitoring is also enabled for HP-UX and Solaris systems.
 
Follow these steps:
 
  1. Navigate to the 
    cdm > 
    Host name
     > Network 
    node.
  2. Specify the 
    Interval (minutes)
     between probe requests for monitoring information.
    Reduce the interval to generate alarms and QoS frequently. A smaller value can also increase the system load.
  3. Select the required monitor from the list and then 
    Publish Data
     to generate QoS messages for the monitor.
  4. Update the following information to configure the thresholds for the alarm message:
    •  
      Enable High Threshold: 
      enables the probe to generate an alarm if the monitored value is greater than the specified value. CA recommends you to specify a higher severity alarm for High threshold.
      •  
        Threshold: 
        specifies the maximum value. The probe generates a higher severity alarm if the monitored value is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the monitored value is greater than the specified value.
        Default severity: Error
    •  
      Enable Low Threshold: 
      enables the probe to generate an alarm if the monitored value is greater than the specified value. CA recommends you to specify a lower severity alarm for Low threshold.
      •  
        Threshold: 
        specifies the maximum value. The probe generates a lower severity alarm if the monitored value is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the monitored value is greater than the specified value.
        Default severity: Warning
  5. Select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  6. Click 
    Save
    .
Configure CPU Monitoring
You can monitor the combined performance of the CPUs in a system and also for each individual CPU in a multi-CPU system. You can configure the number of samples and retrieve the average of computed values. All calculations are based on the number of CPU ticks returned, for example, the /proc/stat command returns the value in Linux. The probe adds the column values (user, nice, system, idle, and iowait) for calculating the total CPU ticks. In a multi-CPU environment, the total for all CPU column values are added. The delta values are also calculated by comparing the total CPU tick values of last and current interval. Then, the percentage values are calculated for each column based on the total CPU ticks value. The QoS for total CPU value is the sum of CPU System, CPU User, and (if configured) CPU Wait.
A multi-CPU (multi-core processor) is a single computing component with two or more independent processors called cores. Each core individually reads and executes program instructions. A multi-core processor implements multiprocessing in a single physical package.
 
(From version 6.10 on AIX only)
 You can also generate QoS messages for the 
Processor > Total CPU > Smt
 to monitor the number of simultaneous multithreads in an LPAR.
 
Follow these steps:
 
  1. Navigate to the 
    cdm > 
    Host name
     > Processor 
    node.
  2. Under the 
    Processor Configuration
     section, update the following information to configure monitoring parameters:
    Reduce the intervals and number of samples to generate alarms and QoS frequently. Smaller values can also increase the system load.
    •  
      Interval (minutes): 
      specifies the interval (in minutes) after which the monitoring information is obtained.
    •  
      Samples: 
      specifies the number of samples the probe stores to calculate values to compare with thresholds.
      : Set the sample value to 0 to use the default sample value for QoS messages.
    •  
      QoS Interval (Multiple of 'Interval')
      : specifies the monitoring interval between each sample collection for QoS information. For example, If the interval is set to 5 minutes and number of samples is set to 5, the average CPU utilization for the last 25 minutes is displayed in the QoS message.
       If “the value is greater than 1, the probe only calculates the average data for CPU monitoring. For disk, memory, and paging QoS messages, the probe generates the QoS messages with the current value. The probe only uses the QoS Interval to reduce the frequency of QoS messages.
  3. Select 
    Set QoS Target as 'Total' 
    to specify the source of the QoS for 
    Total (Individual as well as Average)
     from the hostname to 
    Total
    . The following SQL scripts demonstrate how to update old data to confirm with when the QoS Target as 
    Total
     is changed:
    •  
      QOS_CPU_USAGE
      Execute the following SQL query in the UIM database to view the rows to be updated:
      SELECT * FROM dbo.s_qos_data
      WHERE probe LIKE 'cdm'
      AND qos LIKE 'qos_cpu_usage'
      AND target NOT IN('user','system','wait','idle')
      Execute the following SQL query in the UIM database to update the table for the new target. 
      Target
       is the new QoS target to be set and 
      Source
       is the QoS source for which target is changed. You can configure both the values, as applicable.
      Declare @Target varchar(100) Declare @Source varchar(100)
      SELECT @Target = 'Total'
      SELECT @Source = 'tsuse10-32'
      UPDATE dbo.s_qos_data
      SET target=@Target
      WHERE source LIKE @Source
      AND probe LIKE 'cdm'
      AND qos LIKE 'qos_cpu_usage'
      AND target NOT IN('user','system','wait','idle')
    •  
      QOS_CPU_MULTI_USAGE
      Execute the following SQL query in the UIM database to view the rows to be updated:
      SELECT * FROM dbo.s_qos_data
      WHERE probe LIKE 'cdm'
      AND qos LIKE 'qos_cpu_multi_usage'
      AND (target NOT LIKE 'User%'
      AND target NOT LIKE 'System%'
      AND target NOT LIKE 'Wait%'
      AND target NOT LIKE 'Idle%')
      Execute the following SQL query in the UIM database to update the table for the new target. 
      Target
       is the new QoS target to be set and 
      Source
       is the QoS source for which target is changed. You can configure both the values, as applicable.
      Declare @Target varchar(100) Declare @Source varchar(100)
      SELECT @Target = 'Total'
      SELECT @Source = 'tsuse10-32'
      UPDATE dbo.s_qos_data
      SET target=@Target+RIGHT(target,2)
      WHERE source LIKE @Source
      AND probe LIKE 'cdm'
      AND qos IN ('qos_cpu_multi_usage')
      AND (target NOT LIKE 'User%'
      AND target NOT LIKE 'System%'
      AND target NOT LIKE 'Wait%'
      AND target NOT LIKE 'Idle%')
  4. Update the following information to configure advanced CPU monitoring parameters:
    •  
      Include CPU Wait in CPU Usage: 
      enables you to add the CPU wait time to the total CPU usage of all and individual CPUs.
    •  
      Number of CPUs:
       displays the number of CPU cores on the host system.
    •  
      Top CPU consuming processes in alarms
      : specifies the number of top CPU consuming processes that are included in CPU usage alarms.
      Default: 5Consider the following points while using this option:
      • This alarm is generated when the defined total CPU usage is breached. The new alarms generate the process information in the following format:
        [processname[pid]-cpu%]; [processname[pid]-cpu%]
      • The actual CPU value in the alarm may not always match the total percentage of all the top CPU consuming processes shown in the alarm message. It may vary as Total CPU Usage is calculated on the basis of samples. The probe retrieves the raw data at a given time and displays in the alarm.
      • For non-Windows platform, the probe uses 
        ps
         command to retrieve the top CPU consuming processes.
      • Depending on your environment, the values achieved (%) can be over 100%. For example, the monitoring environment includes the 56-core system using hyper-threading. In this case, the number of virtual cores, because of hyper-threading (more than 1 thread per core), becomes twice that value (2 threads per core). If the user has up to 112 virtual cores (112 = 56*2), the maximum possible value (%) from 
        Top CPU consuming processes
         can be 11200% (112 virtual cores * 100). This explains why a user may see CPU consuming values over 100% or 1000% or 5000%.
  5.  
    (Windows only) 
    Select the 
    Processor Queue Length
     monitor and select 
    Publish Data
     and 
    Publish Alarms
     fields to retrieve QoS data and generate alarm messages, respectively. The QoS message displays the number of queued processes divided by the number of processors waiting for CPU time for the system.
  6.  
    (Windows only) 
    Update the following information to configure CPU queue length monitoring parameters:
    •  
      Maximum Queue Length
      : specifies the maximum number of queued processes.
    •  
      Alarm Message:
       specifies the alarm message to generate if the monitored value is greater than the specified value.
  7.  
    (Linux, Solaris, AIX and HP-UX platforms) 
    Select the required monitor and select 
    Publish Data
     and 
    Publish Alarms
     fields to retrieve QoS data and generate alarm messages, respectively. You can select from the following monitors:
    •  
      Load Average 1 min: 
      monitors the average CPU usage during the last one minute
      .
       
    •  
      Load Average 5 min: 
      monitors the average CPU usage during the last five minutes.
    •  
      Load Average 15 min: 
      monitors the average CPU usage during the last fifteen minutes.
    •  
      System Load:
       monitors the CPU load on the monitored system.
     The load average monitors are only available if 
    Calculate Load Average per Processor 
    is selected in the 
    cdm
     node.
  8.  
    (Linux, Solaris, AIX and HP-UX platforms) 
    Update the following information to configure load average monitoring parameters:
    •  
      (Load Average monitors) Max. Load Average
      : specifies the maximum average load on the CPU in the monitored duration.
    •  
      (System Load monitor) Maximum System Load
      : specifies the maximum load on the CPU.
    •  
      Alarm Message:
       specifies the alarm message to generate if the average CPU load is greater than the specified value.
  9. Under the 
    Individual CPU
     node 
    > Monitors
     section, select from the following monitors:
    •  
      CPU Usage Difference:
       monitors the maximum difference in usage between any two CPUs in a multi-CPU system. The probe generates an alarm if the difference is greater than the specified value.
    •  
      Maximum CPU Usage:
       monitors the the maximum usage of any CPU in a multi-CPU system. The probe generates an alarm if the CPU usage is greater than the specified value.
  10. Update the following information to configure alarms:
    •  
      Publish Alarms: 
      enables the probe to generate an alarm if the monitored value is greater than the specified value.
      •  
        Threshold: 
        specifies the maximum value. The probe generates the alarm if the monitored value is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the monitored value is greater than the specified value.
  11. Select the required monitor from the list and then 
    Publish Data
     to generate QoS messages for the monitor.
  12. Select 
    Compute Baseline
     to enable thresholds. This option might not be available depending on your CA Unified Infrastructure Management configuration. For more information, see Configuring Alarm Thresholds.
  13. Click 
    Save
    .
(Linux, Solaris, and AIX platforms) Configure IOStat Monitoring
You can monitor the input and output statistics (iostat) of the monitored devices. The probe executes the iostat command to retrieve the iostat monitoring values. The QoS values are obtained from the second sample value collected from the monitored devices.
 Consider the following information before the configuration:
  • Ensure that 
    Monitor iostat
     is selected in the 
    General Configuration
     section of the 
    cdm
     node.
  • Ensure that the necessary package to run the iostat command is present in the system where the cdm probe is deployed.
 
Follow these steps:
 
  1. Navigate to the 
    iostat
     node.
  2. Update the following information to configure the iostat monitoring properties:
    •  
      Interval (minutes)
      : specifies the time interval (in minutes) after which the monitoring information is obtained.
      Default: 5
    •  
      Sample Interval (seconds)
      : specifies the time interval (in seconds) to retrieve the sample interval values. Specify a lesser value than the 
      Interval
       field.
      Default: 10
    •  
      Ignore Iostat Devices
      : specifies the iostat devices to be excluded from monitoring. For example, specifying the regular expression sda|sdb in this field excludes the sda and sdb iostat devices from monitoring. 
  3. Navigate to the required 
     
    Device name
     
     node. The probe automatically detects the OS and displays supported monitors for the detected OS. For more information about the specific iostat monitors for your OS, see cdm Metrics.
  4.  
    (From version 6.10)
     Select 
    Iostat Disk Missing
     to enable the probe to generate alarms for a missing disk.
  5. Select the required monitor from the list and then 
    Publish Data
     to generate QoS messages for the monitor.
  6. Update the following information to configure the thresholds for the alarm message:
    •  
      Enable High Threshold: 
      enables the probe to generate an alarm if the monitored value is greater than the specified value. CA recommends you to specify a higher severity alarm for High threshold.
      •  
        Threshold: 
        specifies the maximum value. The probe generates a higher severity alarm if the monitored value is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the monitored value is greater than the specified value.
        Default severity: Error
    •  
      Enable Low Threshold: 
      enables the probe to generate an alarm if the monitored value is greater than the specified value. CA recommends you to specify a lower severity alarm for Low threshold.
      •  
        Threshold: 
        specifies the maximum value. The probe generates a lower severity alarm if the monitored value is greater than the specified value.
      •  
        Alarm Message:
         specifies the alarm message to generate if the monitored value is greater than the specified value.
        Default severity: Warning
  7.  
    (From version 6.20) 
    Enable or disable the following options to manage the read and write requests for the Iostat messages:
    •  
      Iostat Reads Per Second: 
      enables monitoring of the read requests issued for the device each second.
    •  
      Iostat Writes Per Second: 
      enables monitoring of the write requests issued for the device each second.
    •  
      Iostat Reads Writes Per Second: 
      enables monitoring of both the read and write requests issued for the device each second.
  8. Click 
    Save
    .
View Alarm Messages
You can view alarm messages in the probe.
 
Follow these steps
:
  1. Navigate to the 
    cdm
     node.
  2. Under the 
    Messages
     section, select the required message.
    You can view the following information to configure the message properties:
    •  
      Message ID:
       indicates the unique ID of the alarm message.
       
    •  
      Message Text:
       indicates the text for the alarm message. The probe uses $ signs to include variables.
    •  
      Severity:
       indicates the severity of the alarm.
    •  
      Message Token:
       indicates the category of the alarm message to be identified in USM.
    •  
      Subsystem:
       indicates the subsystem ID of the message.
       CA does not recommend you to assign the same subsystem ID to different messages
      .
       
    •  
      i18n Token:
       indicates the category of the alarm message to be identified in USM in non-english locales.
Configure Advanced Properties
You can configure the following advanced properties of the probe:
  • Configure probe defaults
  • Copy probe configuration parameters
  • Edit the configuration file using raw configure
  • Use regular expressions
For advanced cdm probe configuration and to get answers to frequently asked questions, see cdm Advanced Configuration.